[Paper Review] Upgrade of the MAGIC telescopes

D. Mazin, D. Tescaro+3 more|arXiv (Cornell University)|Oct 19, 2014

Astrophysics and Cosmic Phenomena1 references3 citations

TL;DR

This paper details the 2011–2012 major upgrade of the MAGIC Cherenkov telescopes on La Palma, unifying the readout electronics, trigger systems, and camera designs of both telescopes using DRS4-based readout and upgraded receiver boards. The key result was a stabilized, more sensitive, and robust system with lower dead time, improved dynamic range, and enhanced operational reliability for 5–7 years of sustained high-energy gamma-ray observations.

ABSTRACT

The MAGIC telescopes are two Imaging Atmospheric Cherenkov Telescopes (IACTs) located on the Canary island of La Palma. With 17m diameter mirror dishes and ultra-fast electronics, they provide an energy threshold as low as 50 GeV for observations at low zenith angles. The first MAGIC telescope was taken in operation in 2004 whereas the second one joined in 2009. In 2011 we started a major upgrade program to improve and to unify the stereoscopic system of the two similar but at that time different telescopes. Here we report on the upgrade of the readout electronics and digital trigger of the two telescopes, the upgrade of the camera of the MAGIC I telescope as well as the commissioning of the system after this major upgrade.

Motivation & Objective

• Address performance inconsistencies and aging hardware issues in the original MAGIC I telescope, which used outdated MUX-FADC readout and aging receiver boards with slow threshold control.
• Unify the readout and trigger systems of MAGIC I and MAGIC II to simplify maintenance and improve system robustness.
• Enhance sensitivity and reduce dead time by replacing the camera of MAGIC I with a 1039-pixel, all-small-pixel design identical to MAGIC II.
• Improve real-time control of discriminator thresholds via fast IPRC (individual pixel rate control) using FPGA-based receiver boards.
• Ensure stable, low-noise operation across the entire system through comprehensive calibration of trigger thresholds and signal processing.

Proposed method

• Replaced the MUX-FADC readout system in MAGIC I with a DRS4-based system, sampling at 2 Gsamples/s for improved linearity, lower noise, and reduced dead time (<1%).
• Upgraded the receiver boards of both telescopes to include FPGAs for fast IPRC response (seconds instead of minutes), enabling real-time adjustment of discriminator thresholds.
• Replaced the MAGIC I camera with a clone of the MAGIC II camera, increasing pixel count from 577 to 1039, all 1-inch pixels, to unify design and improve trigger efficiency.
• Replaced the L0 trigger discriminator system with a digital, FPGA-controlled implementation for consistent and fast threshold calibration across all channels.
• Implemented a two-stage calibration process: rate scans to determine optimal discriminator thresholds (set at ~4.5 phe), and delay calibration to optimize L1 coincidence trigger timing.
• Used short, calibrated pulses (FWHM < 2 ns) to calibrate discriminator thresholds in photoelectrons (phe), achieving flat sensitivity across channels.

Experimental results

Research questions

• RQ1 How can the performance and reliability of the aging MAGIC I telescope be improved to match the more modern MAGIC II system?
• RQ2 What is the optimal configuration of the L0 and L1 trigger systems to minimize dead time and maximize sensitivity while maintaining low noise?
• RQ3 Can a unified DRS4-based readout system be successfully deployed across both telescopes to reduce cost, complexity, and maintenance burden?
• RQ4 How can real-time individual pixel rate control (IPRC) be implemented to maintain stable trigger rates under varying sky conditions?
• RQ5 What calibration procedures ensure flat sensitivity and consistent performance across all 1039 trigger channels after the upgrade?

Key findings

• The upgraded system achieved a dead time of less than 1% across all channels, significantly improving data acquisition efficiency.
• The L1 trigger gate width was measured at (6 ± 0.5) ns, with L0 trigger FWHM set to 5.5 ns to ensure consistent and stable signal acceptance.
• Discriminator thresholds were calibrated to 4.5 phe, resulting in a stereo trigger rate of ~280 Hz, with only ~40 Hz due to chance coincidences.
• The IPRC system reduced threshold adjustment time from minutes to seconds, enabling real-time response to changing sky conditions and NSB levels.
• The upgraded MAGIC I camera with 1039 pixels and the unified DRS4 readout system achieved a 1.66× larger effective trigger area compared to the original MAGIC I.
• The commissioning process confirmed stable performance over multiple nights, with rate scans showing consistent trigger behavior and no significant drift in calibration.